TL: EXPERIMENTAL PREDATOR CONTROL MEASURES ON MARINE SALMON FARMS IN SHETLAND SO: Holly Arnold, Greenpeace UK (GP) DT: January 1992 Keywords: oceans fisheries fish salmon uk europe greenpeace reports aquaculture gp / A Report by Holly Arnold for Greenpeace UK Submission to the Planning and Coordinating Committee of the Marine Mammal Action Plan, United Nations Environment Programme January, 1992 SUMMARY The exposed rocky shores and sheltered sea lochs of Shetland are the natural habitat for significant populations of grey and common seals. Over sixty salmon farms are now sited around the Shetland coast, leading to a degree of conflict with indigenous predators, whose varied diet includes the species of wild fish attracted to the farm sites, as well as the occasional salmon. The activities of seals around salmon farms can result in large economic losses to this relatively new industry. Estimates of the losses incurred by the Scottish industry from damage claimed to be inflicted by predators range from œ1,400,000 to œ4,800,000 per year (Ross, 1988). Some salmon farmers shoot seals in order to reduce predation on their sites. This report describes and evaluates alternative, non-destructive methods of eliminating the impact of seal predation on salmon farms, with emphasis on a relatively new system of "cage net tensioning" which has been developed in Shetland. The most widely used systems for deterring predators on salmon farms are underwater predator nets, which consist of various arrangements of nets hung outside of the salmon stock cages to form a barrier between the salmon and potential predators. The main types in use are box or envelope nets which enclose the entire cage, and curtain nets which are deployed in panels around the salmon pens, screening the stock from access by seals. However, strong currents can cause billowing, pushing the outer exclusion net and the inner stock net together, thus negating their effectiveness and allowing predators access to the fish. Seals and diving birds become entangled in these nets and drown, and exclusion nets also represent an additional expense in the farmer's equipment and maintenance budget. Seal scarers, which emit random sounds at a frequency distressing to seals, show some promise as deterrents in the short term but have proved unreliable when used regularly as seals quickly become habituated to the disturbance. During the past few years, some Shetland salmon farmers have designed and tested several methods of increasing the rigidity of the fish cage to prevent seals pushing the netting inwards in order to catch the salmon, thus precluding the need for predator nets and the extra expense and maintenance that they entail. These "net tensioning" systems, currently employed by at least nine salmon farming companies in Shetland, are inexpensive to install and simple to make and use. To date, the various tensioning systems appear to have reduced seal predation at a number of farm sites. Apparently, since installing the systems, none of the operators have felt it necessary to shoot seals at their farms. Few quantitative data are available on the scale of predation or stock losses on salmon farms, preventing a full evaluation of anti-predator devices. Although one estimate suggests that approximately 1,050 seals may be killed annually on Scottish salmon farms (Ross, 1988), the true number of seals shot or killed by entanglement each year is unknown. However, net tensioning systems offer an alternative to other less effective predator control methods currently employed on UK salmon farms, and in particular the unnecessary shooting of seals. CONTENTS 1. INTRODUCTION 2. SEAL PREDATION ON SALMON FARMS 2.1 Seal Populations in Scotland 2.2 Legislation Affecting Seals in the UK 2.3 Salmon Farm Losses From Seals 3. ATTITUDES TO SEAL PREDATION AND PREDATOR CONTROL 3.1 Salmon Farmers' Associations 3.2 Salmon Farmers 4. PREDATOR CONTROL METHODS 4.1 Cage Net Design 4.2 Underwater Predator Nets 4.2.1 Curtain Nets 4.2.2 Box Nets 4.2.3 Bag/Envelope Nets 4.2.4 Effectiveness of Predator Nets. 4.3 Acoustic Seal Scarers 4.4 Seal Shooting 4.5 Other Methods 4.6 Current Research 5. CAGE NET TENSIONING SYSTEMS 5.1 Sinker Tubes or Pipe Frames 5.1.1 Rigid Joints 5.1.2 Flexible Joints 5.1.3 Disadvantages 5.2 Farm Ocean Cages 5.3 Cage Net Tensioning with Weights 5.3.1 Individual Pulley Weights 5.3.2 Linked Tensioning with Pulley Weights 5.3.3 Linked Tensioning with Slip Loops 5.4 Advantages and Disadvantages of Cage Net Tensioning Systems 5.5 Effectiveness of Cage Tensioning in Reducing Seal Predation 6. CONCLUSIONS 7. REFERENCES FIGURES (omitted .. unscannable) 1. Shetland Salmon Farms and Seal Distribution 2. Underwater Predator Nets 3. Sinker Tube/Frame Cage Tensioning Systems 4. Farm Ocean Cage 5. Individual Pulley Weight Cage Tensioning 6. Linked Cage Tensioning With Pulley Weights 7. Linked Cage Tensioning With Slip Loops TABLES (omitted .. unscannable) 1. Cage Net Tensioning Devices 1. INTRODUCTION Marine farming of the Atlantic Salmon (Salmo salar) in the Scottish Highlands and Islands began on isolated sites about twenty years ago. During the past decade it has been the boom industry in Scotland, creating jobs for some 6,300 people employed directly on the salmon farms or in supporting industries (Northern Corporate Communications, 1991). There are currently 220 marine salmon farms in Scotland and the Outer Isles, including 60 farms and 7 hatcheries in Shetland (Highlands and Islands Enterprise, 1991). Salmon farming started in Shetland in 1982 and took off in 1984 with an annual production of 50 tonnes; production expanded rapidly and is estimated to reach approximately 9,000 tonnes in 1991, a 180- fold increase over the past seven years. After oil, salmon farming is the second largest industry in Shetland, producing approximately one third of the total U.K. farmed salmon production (C. Young, pers. comm.). The deep, sheltered sea lochs fringing the coast of Scotland and its Islands, and the secluded lochs of the Shetland coast, are natural habitats for large populations of fish-eating wildlife including birds, seals and otters. Figure I shows the distribution of salmon farms and seal colonies around the Shetland coast. The potential for conflict is thus created between predatory species attracted to a concentrated food source and salmon farmers who employ a variety of measures to protect their fish stocks, including the disturbance and killing of wildlife. Killing natural predators is not an acceptable or effective solution to the problem of seal predation on salmon farms. At the third meeting of the Planning and Co-Ordinating Committee (PCC) of the UNEP Marine Mammal Action Plan held in January, 1991, it was brought to the attention of the delegates that a salmon farmer in Shetland has devised a method of tensioning his nets which has apparently eliminated seal predation and mortality on his farms. The PCC requested that this matter be studied further and a report presented to the next committee meeting (Anon,1991). This report describes "net-tensioning devices" in detail and the alternative methods operators are employing to reduce seal predation on Shetland salmon farms. Research was conducted in Shetland where net tensioning systems were designed, and during one week's visit in October, 1991, owners or site managers of twenty three salmon farming companies were contacted. Between them these companies run 34 farm sites. Nine farm sites were visited; interviews or brief discussions were held either in person or over the telephone, to determine the extent of their predator problems and the methods of control or deterrent employed. Those farmers using net tensioning systems were sought to obtain details of the design, management, and efficiency of such measures in reducing seal predation. The following represents a brief overview. For a broader study of predation and control measures on Scottish salmon farms see Ross (1988), Campbell (1991) and Howell and Munford (in press). Because the interviews with salmon farmers were not uniform in either the amount or detail of information collected, no quantitative analysis of the responses was possible. This information must therefore be regarded solely as qualitative. 2. SEAL PREDATION ON SALMON FARMS A detailed survey conducted by the Marine Conservation Society in 1988 included information acquired by questionnaires and interviews from approximately one third of the fish farms in Scotland (Ross, 1988). Over 80% of the farms surveyed in 1988, and in the present study, claimed to suffer damage from seals and the recent brief survey in Shetland confirmed that seals are considered the major predators capable of causing greatest damage to salmon stocks. 2.1 SEAL POPULATIONS IN SCOTLAND The two species of seal commonly found in Great Britain are the grey seal (Halichoerus grypus) and the common or harbour seal (Phoca vitulina). Populations of both species are widely distributed along the west coast of Scotland, its Outer Islands and around the coast of Shetland. Grey seals are also found on the east coast of Scotland as are isolated colonies of common seals. The most recent estimate of the total grey seal population in Scotland for 1991 is between 79-85,000. The total estimated minimum common seal population in Scotland is 19-20,000 (Harwood, et al., 1991). The Scottish common and grey seal populations appear to have suffered minimally from the phocine distemper virus epizootic in 1988 (Harwood, et al, 1991). A survey of common seals carried out in Shetland in August of 1991 indicates that the population there is an estimated 4,700 (C. Duck, pers.comm). The previous survey in 1984 gave a similar estimate, which suggests that the population is probably stable at this level. The current grey seal population in Shetland is estimated to be 3,500 (Harwood, et al, 1991) . 2.2. LEGISLATION AFFECTING SEALS IN THE UK Both species of seals found in the UK are protected by the Conservation of Seals Act 1970 which prohibits the taking, injuring or killing of seals during the close season and also lays down the conditions and methods under which seals may be killed. The aim is to protect the animals during their breeding seasons. As a result of the phocine distemper virus in 1988, which led to the death of large numbers of seals (especially common seals) in European waters, a decision was taken to extend the protection of seals. The Conservation of Seals (England and Wales) Order 1988 and the Conservation of Seals (Common Seals) Order 1988 were introduced. Both of these orders had the effect of extending the close season to the whole year for both species in England and Wales and for common seals only in Scotland. This protection was extended for a further 3 years on the east coast of England from December 1990 for both species of seals (the Conservation of Seals (England) Order 1990) while in Scotland protection was continued indefinitely for common seals in the Shetland area through the Conservation of Seals (Common Seals) (Shetland Islands Area) Order of 1990 and 1991. Licences can still be issued for predatory control and research purposes during the close season while the shooting of seals in the vicinity of, and damaging fishing gear may still be carried out without the need for a licence. There is also a requirement to report the number of seals shot under licence on a monthly basis. In 1989 and 1990, no licences were granted to shoot either species of seal in Scotland, and one was issued in 1991. In the Shetland Islands, an application to establish a salmon farm must be made to the Shetland Islands Council; licences are granted after consultation with the NCCS and other authorities, and include a requirement that farmers use predator nets or other predator control measures. 2.3 SALMON FARM LOSSES FROM SEALS Recent research in Scotland, England and Wales into grey seal diet suggests that sand eels and gadoids (cod, whiting, saithe) comprise 78-97% of their natural food (Hammond & Prime, 1990). Studies of winter foraging by common seals in Scotland have found that sprat and herring form over 90% of their diet (Thompson, et al., 1991); in neither of these studies of wild seal populations was salmon found to be part of the diet. These results have changed the perception gained by research on seal diet in the 1970's. As these earlier studies examined stomachs of seals shot primarily around salmon netting stations, so it is not surprising that they concluded that salmon was an important part of the seal diet. Although primarily fish eaters, seals are opportunistic feeders and probably prey upon whatever food is available in the areas they frequent. It is therefore no surprise that seals are attracted to salmon farms where large volumes of fish are concentrated into and around relatively small enclosures, presenting a potential take- away meal. Other wild fish including saithe (Pollachius virens) swim near the outside of the cages in their thousands. Attracted by the food pellets and waste as well as the organisms (mainly algae, tunicates and molluscs) fouling the nets, these fish are exploited by predators. The scale and frequency of seal predation is unpredictable and unique to each site. Farm operators agree that the most common mode of seal attack is from beneath the cage, although predators will approach from the net sides as well. The animals apparently charge at the floor of the cage from below. This drives the net against the cage where fish can be slashed, bitten, grabbed or sucked through the net (See also Ross, 1988.). Farmers have claimed that, following a seal attack, some of the salmon which are not physically damaged may become stressed by shock and disturbance with a subsequent reduction in feeding for a week or more. In addition, predators may damage stock cages to gain entry, occasionally releasing large numbers of fish through holes torn or chewed through the nets. Similar observations on the effects of seal predation were also noted in previous surveys of Scottish fish farms (Ross, 1988; Campbell, 1991). Some farmers observed that any increase in activity on the cage units, such as changing nets, harvesting or grading often seems to provoke seals' curiosity and the animals appear more frequently close to the cages. Newly cleaned predator nets weigh less, offering less resistance when pushed, and no longer obscure the fish inside from view, thus possibly restimulating predator interest. Some sites do not employ anti-predator devices and have no problem with seal predation. The majority of the farmers contacted in Shetland said that seals were seen frequently near their cage sites but described their presence as usually just an irritant. Attacks were commonly attributed to the occasional 'rogue' seal, often thought to be a young bull who might take small numbers of fish over a period of weeks, or indulge in an apparent feeding frenzy, killing, damaging or releasing large numbers of fish in one night. The view that such attacks are attributable to 'rogue' seals appears to be supported by behavioural observations (P. Thomson, as cited in Howell and Munford, in press). Seal attacks can result in a range of losses from one or two fish to œ1,000's worth of salmon in one night (Ross, 1988; Campbell, 1991). The frequency of seal problems did not necessarily bear any relationship to the numbers of seals reported to be nearby. 3. ATTITUDES TO SEAL PREDATION AND PREDATOR CONTROL 3.1 SALMON FARMERS' ASSOCIATIONS The Shetland Salmon Farmers Association (SSFA) was formed in 1984 by local salmon farmers as a body to regulate, represent and advise the industry in Shetland, to ensure good communication among the farmers themselves and between the industry and interested government and conservation bodies. The Association has worked closely with local representatives of the Nature Conservancy Council for Scotland (NCCS) since its inception and has maintained a good relationship with conservation non-governmental organisations (NGOs). In the rest of Scotland, the industry is represented by the Scottish Salmon Growers Association (SSGA) which has also issued predator control guidelines to its members, prepared in conjunction with a number of NGOs and government bodies (SSGA, 1990). The SSFA formed an Environmental Working Group in 1988 which includes representatives of the NCCS, and in 1989 issued anti- predator guidelines which were circulated to their members. These give clear details of recommended predator control measures and stress the importance of using such methods conscientiously from the start, before predator problems develop and "with the minimum environmental impact on the predators themselves" (SSFA, 1989). Seal shooting is considered by the Association as a last resort when all other passive deterrents have failed and the SSFA and Scottish Salmon Growers Association guidelines stress this emphatically. In January, 1991 a questionnaire was sent to all SSFA members, requesting information on the types of predator nets and other control measures employed on each farm, as well as details of seals entangled or shot at each site during the past year. This information has been made available to the NCCS in confidence, contributing quantitative data on the value of various control measures in relation to predator damage and mortality. Farming members of the SSFA Environmental Working Group have been in the forefront of developing more effective anti-predator netting systems and have devised methods of tensioning the salmon cages, rather than using predator nets, to deter predators. Following trials of cage net tensioning over the past year and more, details of the successful methods were circulated to members of the Association in June, 1991. 3.2 SALMON FARMERS The farmers interviewed preferred not to kill seals as it is difficult to locate and shoot such a mobile target; hunting down the individual causing damage may take days. Most operators expressed the view that they would rather live with the local wildlife than kill it. They seemed tolerant of seals and other predators, considering them mainly an irritant except for the odd individual attacks which varied in frequency and severity. Large-scale predator damage appeared to be a problem at only a few farms. Seal colonies are valued by some as a natural part of their habitat, and the majority of farmers showed an appreciation of the relatively undisturbed environment in Shetland. Salmon farmers are also aware that blatant incidents of seal shooting can harm their market. 4. PREDATOR CONTROL METHODS 4.1 CAGE NET DESIGN The netted enclosures in which farmed salmon are held are referred to as pens or cages. The original cages used in the early years of salmon farming were made of rigid wire mesh panels but their use was discontinued for reasons of economy and maintenance and replaced by nylon mesh containers. A typical marine salmon cage consists of a cuboid or cylindrical net bag, open at the top and suspended from a flotation collar. A railed walkway provides safe access to the pens as well as a framework onto which a stock pen is hung at least a metre clear of the water to prevent fish jumping out of the cage. A variety of cage shapes, designs and sizes are used throughout the salmon industry, some of which are available commercially and others 'homemade' to the farmers' specifications. Tidal regime, water circulation, extent of fouling and predation are unique to each site. As a result, many farms use a combination of cage shapes and sizes depending on the conditions at each site and the grades of fish in stock. The nylon cage nets usually have a mesh size of 10-25mm, depending upon the size of the fish, and are reinforced with rope along the top, side and bottom edges of the enclosure. Salmon pen nets hang to a depth of 5-15m underwater, varying with the depth of the site and the size of the cage flotation collar. The net is weighted at the corners and base with leaded ropes and/or weights to maintain its optimum shape and volume in variable tidal and current conditions. Some square cage nets are designed with loops at one metre intervals along the edges of the wall and floor joins for the suspension of further weights to ensure that the cage is held uniformly taut. The cylindrical "Polar-cirkel" cage net has reinforced rising ropes stretching from the top to the floor of the net at intervals round the circumference, with integral loops at the bottom of the risers for the attachment of weights. Fouling of nets by algae and encrusting invertebrates is a recurrent problem, particularly during the spring and summer. Nets are commonly treated with a red copper-based antifoulant, which hardens the fibres. In combination with the use of weights, this can help to deter predators. At most sites, nets are lifted at least once a year for cleaning and repair. Most of the farmers contacted use top nets or lines, which are fixed to the handrails about 11.5m above the water surface of each cage, covering the open top of the pen to deter bird predation. 4.2 UNDERWATER PREDATOR NETS A predator net totally surrounds or screens one or more sides of the stock net underwater, thus forming a physical barrier between underwater predators and the fish cage. Once the nets become fouled, they also provide an element of visual screening and the extra weight may provide more rigidity. Predator nets are hung from the outside of the walkway or outer ring of the flotation collar to separate them from the inner cage. The recommended distance from the stock net is at least one metre (SSFA, 1989). The mesh size is commonly 75-150mm. Young seals have been known to become entangled in nets of 100mm mesh and over, and the SSFA recommends that nothing larger than a 75mm mesh be used on all exclusion nets to reduce the possibility of entangling wildlife. Loops at the bottom corners and sides, as well as reinforced seams or ropes running from top to bottom provide for the attachment of weights. 4.2.1 Curtain Nets Curtain or skirt nets consist of single or overlapping panels of net hung vertically round the outside of a single stock cage, or raft of cages, to a depth of five to ten metres or more beneath the floor of the pen (See Fig.2a). Some farmers anchor the curtain nets on the seabed. Seals may persistently attack a cage in one area of the net and a single curtain panel can be hung across the most vulnerable area when necessary. The recent SSFA survey indicates that curtain nets are used by eight salmon farming companies either solely or in addition to other predator nets. One of the main disadvantages of curtain nets is that diving predators may swim underneath the panels to reach the floor and sides of the cage, with the risk to the farmer of stock damage, and to the animal of becoming trapped between the fish pen and the predator net. Also, water currents and tidal fluctuations can billow the net panels, pushing them against the main pen and eliminating any barrier effect. Slack nets can increase the probability of entangling seals and/or diving birds such as cormorants, shags and guillemots. One farmer estimated that some twenty shags per month were caught in the curtain nets formerly used at his site. As curtain nets are considered a poorer option than other exclusion nets, as well as having a high entanglement rate, the SSFA guidelines recommend they be phased out as soon as practicable. However, they are recognised by the SSGA as easier to deploy and maintain, and may give sufficient protection in some circumstances. Only one of the operators contacted in Shetland used curtain nets as the sole form of predator control and these are being phased out and replaced with cage net tensioning. Several farms retain curtain panels in store for use in the event of recurring predator attacks from the same front. 4.2.2 Box or Envelope Nets The box or envelope net forms a complete box round a single pen or a raft of two or more cages (See Fig.2b). As the box net envelopes the entire stock net, it is also referred to as an envelope net. The bottom panel of the box net should hang at least two metres below the cage floor, and will frequently be more in practice due to sagging under its own weight. Weights hung from the bottom corners and sides keep the net taut and clear of the inner fish pen. Some operators also hang weights from the center of the bottom panel to increase its rigidity and resistance to being pushed up from below. Ten of the operators interviewed in October use box nets at thirteen sites on square or hexagonal cages. This type of net is used by 38 of the 50 respond ants to the SSFA survey; they are also the most commonly deployed exclusion nets in Scotland (Ross, 1988). Box nets are considered more effective than curtain nets by both the Scottish and Shetland salmon farmers' associations (SSGA, 1990; SSFA, 1989).The SSFA recommends them as the preferred predator nets and stresses the need to keep them well maintained and taut. However, box nets are rendered ineffective by rough weather, strong currents, and high tides which can billow and deform both the inner and outer nets. Areas slackened by deformation increase the likelihood of entangling wildlife and most operators mentioned that drowned diving birds trapped in the predator nets were a fairly regular occurrence. 4.2.3 Bag Nets Bag net is also an alternate term for box or envelope nets, but usually refers to exclusion nets hung from the outer flotation ring of the round Polar-cirkel cages (see Fig. 2c). Leaded 1 kg per metre rope is sewn around the bottom edge of the cage and bag nets as well as in a cross pattern in the floor of both nets to provide extra weight. Further individual weights are hung from the reinforced rising ropes round each net. The shape of the bag net allows the floor panel to sag, and the distance between the two nets at this point varies from 3-6m. The system as illustrated in Figure 2c represents the most efficient method described for weighting bag nets. Bag nets have disadvantages similar to those mentioned for box predator nets with the additional problem of maintaining sufficient distance from the cage near the water surface. The inner and outer faces of the tubes supporting the nets are less than a meter apart, increasing the likelihood of the nets swinging together near the surface in rough weather, even when well weighted from below. 4.2.4 Effectiveness of Predator Nets As seals are frequently observed in the vicinity of most salmon farms, predator nets are considered by most farmers to be a necessity for protection of their stock and investment. Most of the farmers contacted who use predator nets considered them a sufficient deterrent and did not feel a need for an alternative form of predator control. Box and/or bag nets are used at 18 sites; two of the companies contacted, running two sites each, still use envelope nets at one site and cage net tensioning at the other. Although all but one of the fifty respondents to the SSFA survey use predator nets, Howell and Munford (in press) suggest that it might be prudent to delay investing in anti-predator measures until it is clear that stock is being lost. The scale and nature of predation problems is highly variable between individual sites. However, both the Scottish Growers and Shetland Farmers Associations strongly advise that such measures be considered in the earliest stages of planning and management of a salmon farm (SSFA, 1989; SSGA, 1990). Farm operators were asked to estimate how much they spent annually on predator nets and other control measures and what proportion of their annual budget this represented. Few were able to give specific figures for the expense of maintaining predator controls but the common theme was that their cost was usually insignificant in relation to other expenses. Environmental conditions, inadequate maintenance, and poor husbandry can render predator nets ineffective. There is a need to find a system which separates wildlife from stock effectively without causing entanglement, is adaptable to the extremes of existing conditions, requires less maintenance, and is therefore more acceptable to salmon farmers than predator nets. 4.3 ACOUSTIC SEAL SCARERS Over the years, a number of experiments have been made into the use of sound to scare seals away from fishing nets and aqua culture cages. Weighted firecrackers which explode underwater near the nets, underwater playback of recordings of killer whale vocalizations and other trials with a range of noises used near common seals in the vicinity of a salmon netting station produced no reliable deterrent effect. The seals quickly became accustomed to the negative stimuli and stopped responding to them (Bonner, 1982; Howell and Munford, in press). Seal scarers are now commercially available and work on the principle of emitting bursts of unpleasant high frequency audio signals at random intervals. Transmission frequency is usually around 10 Khz, near the peak of seals' hearing sensitivity. Operators who have used scarers say that they can be mechanically unreliable and do not function for long as a deterrent as the animals become habituated to the noise, and may even be attracted to the site. 4.4 SEAL SHOOTING The shooting of seals is sometimes seen by salmon farmers as a quick and relatively cheap alternative to passive predator controls. Both the SSFA and SSGA guidelines stress the undesirability (and futility) of seal killing and recommend that it be considered only as a last resort when all other deterrent measures fail. However, destruction of individual animals is of negligible benefit to the farmer as seals are highly mobile and "... immigration is likely to replace the animals killed or dispersion will shift the problem elsewhere and permit it to return once the persecution is relaxed" (Howell and Munford, in press). Evidence suggests, however, that shooting is a common predator control practice on salmon farms. In light of a survey of Scottish fish farms in 1988, Ross (1988) estimated that over 1,000 seals (both common and grey) were shot each year. Industry figures, perhaps not surprisingly, tend to give much lower levels. Sources within the industry estimate that the number of seals reported killed on salmon farms in Scotland during the past year is some 3-400, including less than 100 animals killed in Shetland, approximately one fifth of which died as a result of entanglement. The SOAFD, in conjunction with NERC, recently conducted a voluntary survey of the number of seals killed at fish farms and the results are being analysed. The incidence of seal shooting appeared to be minimal among the Shetland salmon farmers interviewed. 4.5 OTHER METHODS Although setting nets with the object of entangling seals is illegal and unacceptable as a method of predator control (SSGA, 1990, SSFA, 1989), a survey of 47 Scottish salmon farms in 1988 revealed that 319 seals were reported killed in a year; 113 deaths (approximately one third of the total killed) were caused by entanglement, which in some cases appeared to be deliberate (Ross, 1988). Details of other predator deterrent measures used on Scottish farms can be found in Ross (1988) and Campbell (1991). 4.6 CURRENT RESEARCH In September 1991, trials were conducted in Scotland by Dr. A.D. Hawkins of the Scottish Office Marine Laboratory with a mechanical acoustic seal scarer being developed by the Yorkshire engineer Roger Flute. This experimental device emits sounds in frequencies from 2-20kh and produced 'spectacular' results among wild seal populations who splash dived or swam away the instant the equipment was activated. Funding is currently being sought for further research into the development of a suitable power source. Representatives of the fish farming industry have been contacted to arrange future trials of a prototype on salmon farms (A.D. Hawkins, pers. comm.). A review and evaluation of seal scaring devices is being conducted in Canada by Phoenix Marine Consulting of British Columbia. Since 1990, the University of Aberdeen Aquaculture Department has been carrying out an environmental audit of salmon farms in Scotland, Orkney and Shetland, for Ghillie & Glen, salmon smokers. The Department has been commissioned to examine all aspects of the salmon farming industry and its effects on wildlife and the environment, including the use of predator controls and their effectiveness. This is a continuing study and should produce valuable data which has so far been unavailable for this industry (L. Laird, pers comm). 5. CAGE NET TENSIONING SYSTEMS None of the exclusion net systems described above provide a completely reliable predator deterrent. Another more efficient method of preventing seal predation and fish loss, which also prevents entanglement, would obviously be beneficial to salmon farmers and wildlife alike. Gibbie Johnson, owner of G. Johnson (Shetland) Ltd., began to design and test alternative predator control systems some three to four years ago. Cage net tensioning is a relatively new predator control measure which is being pioneered in Shetland. The primary aim of these systems is to hold the cage or stock net uniformly taut, so that it presents a "wall" to any underwater predator with no slack areas for entanglement or purchase on the net through which a seal can bite or grab at fish. The main advantage of net tensioning cages is that they do not require predator nets. As there is no surrounding predator net, there is no appreciable difference in water circulation through the stock pens, and no extra nets to lift, clean and maintain. Refinements to the tensioning systems are still being developed and several are currently used by at least ten salmon farming companies in Shetland, some employing more than one method on their sites. 5.1 SINKER TUBES OR PIPE FRAMES This system employs a square frame of 50mm diameter steel pipe attached to the bottom edge of the pen, confining the cage net and maintaining it in a rigid box shape, rather than hanging numerous weights from the cage to keep the net taut. The weighting pipe itself is suspended just below the floor of the cage at a fixed depth and the net attached to the heavy frame by loops at each corner (see Figure 3). 5.1.1 Rigid Joints Early designs of this system were made of angled 50mm diameter pipes which slotted into each other to form a solid frame weight. When raised or lowered, the rigid corners twisted or split unless the whole unit was kept precisely on the horizontal to prevent distortion. This method was tried by at least two operators but has now been abandoned because of its inflexibility and handling difficulty. It was estimated by one farmer who used this system that the frame of 100mm diameter pipe on a 15m2 (15m a side) cage weighed approximately 600kg and cost approximately œ2,000. The operator found the system efficient and the large diameter pipe remained rigid during handling. Although this frame was unwieldy, the operator had no predator problems on his site during the year that the sinker tubes were employed. A system of weight and pulley tensioning (see below) is now used at this farm. 5.1.2 Flexible Joints The problems inherent in the above design led to the development of a flexible method of fastening the pipes together, using three links of heavy steel chain to form the corners of the frame (see Fig. 3a). The end links are either concreted, bolted or shackled into adjacent pipe sections, the middle link remaining free. There is sufficient play in the linked frame to avoid the problems of distortion or breaking of the pipes when the tube unit is handled. The frame is suspended from the walkway by fixed ropes near the ends of each pipe which hold the weight in place at a depth just below the base of the cage net. Some operators tie the vertical side panels to the frame only at the mid-point, while others attach the bottom edges at the corners as well as several points along each side of the net. In most cases, the rigidity and weight of the frame hold the bottom panel sufficiently taut to resist a powerful upward rush by a seal. However, at least one farm operator places a sand filled plastic feed sack in the center of the floor panel, thus weighting it further and drawing the center of the floor downward. The weight for this system made of 50mm diameter pipe for 11 m2 cages works out at 200kg (50kg/side) for the frame and 50kg for the sand, a total weight of 250kg per cage. The price of the frames is estimated to be in the region of œ200 per cage, costing approximately œ180 for the pipes and œ20 for the shackles and other fittings. The above system has been in place since 1988 and the operator has experienced no predator problems since then. Another farmer installed sinker tubes with weight in the bottom of the cage in July, 1990 on 12m2 cages and he has found them "a great improvement" over predator nets. To overcome the difficulty of handling the heavy pipe frames each time the net is raised, one farmer has developed a variation of the system by shackling a pulley to each mid-corner link (see Fig.3b). A rope at least two times the depth of the net in length is tied round each bottom corner loop, then run through the pulley and up to the surface, to be tied to the walkway, a stanchion or the handrail. When the net is raised, this running rope is untied to give full slack and the net is lifted from its top corner ropes, leaving the frame in place suspended below the cage. In this case, the bottom net panel does not have extra weight inside it. Although cages are usually rafted in groups of four and more to a block, each cage has its own sinker tube frame. A frame encompassing two or more cages would be too large, heavy and unwieldy to handle. Variations of sinker tube tensioning are used by at least four salmon farming companies in Shetland on four different sites. 5.1.3 Disadvantages The major disadvantages of sinker tubes are their sheer weight and awkwardness in handling. Although two people can lift a corner of the frame, they require winches or extra manpower to lift with the nets attached. The use of pulleys for lifting the stock pen while leaving the frame suspended below the cage allows the net to be manually hauled up, but winches are still required for installation. The fixed ropes suspending the frame may become entangled with the lifting or net attachment ropes when they are slackened for maintenance unless kept sufficiently far from the corners. Frame users estimate a cost of approximately œ200-œ225 per cage, which is less than the cost of predator nets. This expense must be weighed against the severity of the predator problem and its cost to the farmer. Fouling is not an important aspect of pipe frame efficiency unless the sheer volume of encrusting organisms on nets and pipe becomes so massive that the weight on the frame is sufficient to strain and tear the attached nets. 5.2 FARM OCEAN CAGES The Farm Ocean cage is a semi-submersible hexagonal structure of approximately 20m diameter (Fig. 4), sold commercially from Sweden in kit form for approximately œ100,000. A 100-150mm diameter sectional steel pipe frame fits onto the base of the cage which has an integral winch system from the central platform. Predator nets may be deployed from the outside of the walkway if necessary. The principle of tensioning the stock net is the same as that for the home-made sinker tube frame, with the taut cage net forming a rigid barrier to underwater predators. A central weight is attached to the bottom of the cage to ensure that the floor panel remains taut. A central line is also attached to the bottom of the net which can be raised by winching the centre rope from the upper platform. This allows the pen to be raised or lowered for harvesting or maintenance of the nets. At least one of these cages is in use on Shetland and the farmer is satisfied that the deployment of the cage net has been successful in deterring seals without having to hang predator nets. 5.3 CAGE NET TENSIONING WITH WEIGHTS This system of concrete drum weights and pulleys attached to the bottom corners of the cage nets is an adaptable method of maintaining the net shape and holding it taut. Weights can be set to hang from the corners of every cage, or rigged so that two to four cages are tied off to the one weight, thus tensioning each other. Both of these methods are in use in Shetland. A slightly different system is employed at each site, and the following are three representative arrangements from which others could adapt their own to suit particular requirements. 5.3.1 Individual Pulley Weights Drum-shaped 25 litre (120 kg) concrete weights with an eye bolt and pulley attached are hung from a fixed rope at each of the four corners of individual cages (see Fig.5). The cages therefore remain separately weighted even when rafted together in groups of four or eight pens. Each individual cage is held taut in all directions by the four weights at the corners; any slack can be taken up on the pulley rope at the surface. This system has been in use for two years on rafts of 15m2 cages using 25mm mesh fish pens which were formerly weighted by sinker tubes. The weights cost approximately œ20-œ25 each to make up, i.e. œ100 per cage to set up the system. The operator has had no predator problems since installing sinker tubes or individual pulley weight tensioning. Prior to this seals and diving birds were occasionally entangled. 5.3.2 Linked Tensioning with Pulley Weights Another tensioning system currently in use on a raft of ten 12m2 cages employs a single weight to tighten adjacent cages (see Fig.6). At the junction of four cages, one weight, suspended from the walkway tensions all corners. The cage is made of 12mm smote net with a 1.5m deep panel of 75mm mesh attached to the bottom to form a double floor panel. A 120kg weight is sufficient to tension one or two sides of a 12m2 cage; a 250kg drum of concrete forms the central weight which tensions four adjoining cages. The drums hang some 1.5m below the bottom corners of the cages. Ideally, the angle between the rope fastened to the net corner loop and the tensioning rope should be maintained at 45 degrees. When strain is put on any of the lines, either from intentional tightening or water movement shifting the weight, each cage net is tightened against the other through this linked system. A linked tensioning system of this type is used at two Shetland salmon farms. One farmer feels the incidence of predation and entanglement has been reduced, the other feels there has been no change. Both anticipate using seal scarers on occasions as well. 5.3.3 Linked Tensioning with Slip Loops A linked tensioning system, illustrated in Fig.7, is also deployed at another site on a large block of 20m2 steel cages with stock nets of 15m depth. Each cage has 18mm mesh sides and a single 25mm floor net. A 500kg weight is hung from a central point to a depth approximately 5m below the floor of the net. The weighted drum has one or two steel rod cross pieces fixed to the top with an attachment lug at each end. A single loop attached to the bottom corner of the cage net slides along the outer ropes supporting the drum allowing some leeway for the net to ride up and down with tidal changes while retaining its box shape. The vertical tension on the net is maintained by a 1-2kg/m lead rope along the base of the cage side panels. A rope tied to the bottom corner of each cage runs up to the surface through loops fixed at 1m intervals along the sides of the net. This line is then tied to the handrail,and used for raising the nets. To lift the cage, the ropes at either side of the fixed weight suspension cables are slackened off, then the net is winched up by the side ropes. The operator has employed the above or a similar system at one site for three to four years and has also used pipe frames at another site for three years. The experience of this farm is that weight tensioning systems have proved less expensive and easier to use, as well as providing more effective predator control. Drum weight tensioning systems are employed by at least four companies on six salmon farm sites in Shetland. Another salmon farming company is in the process of installing weights and pulleys on its cages at two sites and intends to have the system in use by mid December 1991. 5.4 ADVANTAGES AND DISADVANTAGES OF CAGE NET TENSIONING SYSTEMS One potential difficulty with the tensioning systems described above may occur when the ropes are slack while the cages are lifted. Loose tensioning lines may twist and seize in the pulleys, but untangling them can be dealt with by divers if necessary. Divers are also required to help set the weights and thread ropes through the pulleys, which adds to the costs of installation, but as most sites either employ contract divers on a regular basis or have their own divers to remove dead fish and inspect the condition of live stock and equipment, any extra expense would be negligible. There are numerous advantages to the tensioning systems once in place, the most noticeable being that predator nets are no longer required. The extra lab our and expense involved in lifting, cleaning and maintaining exclusion nets is thus eliminated. The weights attached to predator nets to keep them taut must be lifted with the net each time it is raised. This highlights another significant advantage of tensioned cage nets in that when they are raised, the weighting systems are not lifted as well but remain independently suspended from the walkways. Removing predator nets should improve the water circulation within the cages. Tensioning of salmon cages also ensures that they retain their design volume which helps eliminate stock crowding, contributing to an improvement in fish growth and health. Tensioning devices appear to be less expensive to install and maintain than exclusion nets, as well as more effective at excluding predators, thus providing a more cost- effective deterrent system. Tensioned cage nets can be cleaned in situ with high pressure jet or vacuum hoses, as the mesh is held sufficiently taut to resist being blown away or sucked up during cleaning. 5.5 EFFECTIVENESS OF CAGE NET TENSIONING IN REDUCING SEAL PREDATION Although none of the Shetland salmon farmers contacted appear to have regular or persistent seal predation problems, all but one of the operators employing the system have estimated that replacing predator nets with cage net tensioning has proved 95-100% effective in reducing predation. Seals are still frequently seen near their cages, but the damage to stock has been negligible. The opinion of those using the systems is that tensioning works. Ten of the twenty three companies contacted have used sinker tube and/or drum weight cage tensioning systems for periods of from one month to two-three years. Only two of the operators have encountered problems so far, including one seal entangled at one site and small numbers of fish still being taken by a seal at another. If tensioning is used in combination with a false cage floor and an effective seal scarer, most farmers felt that seal predation would be reduced or eliminated, even on the most seriously affected sites. However, no predator control method devised to date can be guaranteed 100% effective. Good husbandry, proper maintenance, management and hygiene practices are also important elements of predator control on salmon farms. It is very difficult to evaluate the efficiency of cage net tensioning in reducing seal predation and mortality as there are no 'before and after' figures available for entanglement and direct kills. However, the observation that seals are causing less damage to salmon stocks where net tensioning is employed has eliminated the perceived need to shoot seals at these farms and reduced wildlife mortality through entanglement. 6. CONCLUSIONS In order to evaluate specific predator control methods, it is necessary to independently monitor the scale and frequency of seal damage, predation and mortality at individual sites over a period of several years. This would also involve controlled trials using anti-predator systems at a number of sites providing variations in environmental conditions and wildlife density. In addition, an accurate method of determining losses due to predation is needed. It is essential that systems be developed which will eliminate the perceived need for farmers to shoot seals. Such systems must eliminate seal predation and prove cost-effective in order to appeal to the salmon farming industry. As Howell and Munford (in press) point out, there have been no real attempts to date to evaluate the economic significance of stock losses due to seal predation. In Shetland farmers were asked to estimate the amount they spent per year on predator control, the percentage of production this represented, and the stock losses to seals in one year. Of those that replied most felt that the sum spent on predator controls comprised a small proportion of their budget. However, the SSFA and the Shetland Islands Council are planning a cost/benefit analysis of predator control measures employed by the salmon farming industry in Shetland in 1992. The siting of salmon farms in areas not frequented by seal colonies may go some way toward reducing the potential for conflict between wildlife and farm operators. The collection of data and dissemination of information to all salmon farmers on the effective methods available for passive predator control should also contribute to the reduction of direct or indirect killing of wildlife. The Marine Conservation Society held a series of workshops for Scottish salmon farmers during 1990 for this purpose (Campbell, 1991), but communication within the industry must be continued to keep farmers informed of developments in research into stock and environmental welfare. All of the predator nets commonly used on salmon farms have major disadvantages in that they are frequently rendered ineffective by environmental conditions, or are not sufficiently well deployed to separate predators from prey. Both envelope and curtain nets present a potential hazard to wildlife because they can entangle and drown both seals and diving birds. Predator nets also require additional effort to handle and increase the operators' equipment and maintenance expenses. Acoustic seal scarers are still unreliable as a predator deterrent and entail a fairly large initial outlay. The cage net tensioning methods described in this report are simple, accessible, almost maintenance-free systems which appear to function well in preventing loss of salmon to seals without causing harm to wildlife or the marine environment. The development of such systems is a positive step forward on the part of the salmon farmers in Shetland and merits serious investigation and controlled comparison with other methods employed by the industry. ACKNOWLEDGEMENTS This report would not have been possible without the cooperation of the Shetland salmon farming community. I would particularly like to thank J. Moncrieff, C. Young and F. Draper of the Shetland Salmon Farmers Association for their willing assistance. The salmon farmers were helpful and hospitable and I am very grateful to the following for contributing their time and information to this report: B. Anderson, L. Anderson, R. Anderson, W. Doull, G. Duncan, V. Gray, A. Hunter, M. Jamieson, A. Johnson, D. Johnson, Gibbie Johnson, Gordon Johnson, P. Johnson, M. Mitchell, M. Nicolson, F. Oudie, J. Ratter, L. Ritch, D. Robertson, K. Robertson, M. Robertson, A. Tale, Jim Tait, John Tait, V. Rosso, D. Thompson, C. Young A number of people in the salmon farming industry as well as government and voluntary bodies have freely given their assistance and I thank them all for their help, including: R. Berry, J. Buchanan, R. Campbell, C. Duck, A. Hawkins, D. Howell, Q. Jardine, G. Leslie, L. Laird, E. Stuart, T. Svealv, J. Twelves Last but not least, my thanks to all at Greenpeace for their help in the preparation of this report, particularly to S. Mayer for her advice and W. Wickham for his excellent illustrations. REFERENCES Anderson, S. (1981) "Seals in Shetland Waters", Proceedings of the Royal Society of Edinburgh, BOB, pp. 181-188. Anderson, S. (1990) "Seals", Whittet Books Ltd., London. Anon. (1991) Third Meeting of the Planning and Co-Ordinating Committee of the UNEP Marine Mammal Action Plan, Amsterdam, 14-17 January. Bonner, W.N. (1982) "Seals and Man, a Study of Interactions", University of Washington Press, Seattle, 170pp. Campbell, R. (1991) "Non-Destructive Predator Controls on Fish Farms", A Report for the Marine Conservation Society, Ross-on- Wye, England. Cobham Resource Consultants with Fisheries Development Ltd. (1987) "An Environmental Assessment of Fish Farms", Final Report to Countryside Commission for Scotland, Crown Estate Commissioners, Highlands & Islands Development Board, Scottish Salmon Growers' Association, 100pp. The Crown Estate (1987) "Fish Farming, Guidelines on Siting and Design of Marine Fish Farms in Scotland", 13pp. Hammond, P.S. and J.H. Prime (1990), "The Diet of British Grey Seals, Halichoerus grypus", p. 243-254. In W.D. Bowen (ed.) Population Biology of Sealworm (Pseudoterranova decipiens) in Relation to Its Intermediate and Seal Hosts. Can. Bull. Fish. Aquat. Sci.222. Harwood, J. and Greenwood, J.J.D. (1985) "Competition Between British Grey Seals and Fisheries", in Marine Mammals and Fisheries, ed. Begging ton, J.R., Beverton, R.J.H., Lavigne, D.M., George Allen & Unwin, London pp 153-167. Harwood, J., Hiby, L., Thompson, D. and Ward, A. (1991) "Seal Stocks in Great Britain", NERC News, January, 1991 and July, 1991. Highlands & Islands Enterprise (1991) List of Fish Farms in the HIE Area, Inverness, Scotland. Howell, D.L. and Munford, J.G. (in press), "Predator Control on Finfish Farms", A Report for the Nature Conservancy Council for Scotland. Needham, T. (1988) "Seawater Cage Culture of Salmonids", Salmon and Trout Farming, eta. L. Laird and T. Needham, Ellis Horwood Ltd., Chichester, pp. 117-154. Robertson, Morag (1991), Letter from The Scottish Office Agriculture and Fisheries Department, Edinburgh to Greenpeace. Ross, Alison (1988) "Controlling Nature's Predators on Fish Farms", a Report for the Marine Conservation Society, Ross-on- Wye, England, 96pp. Scottish Salmon Growers Association (1990) "Salmon Farming and Predatory Wildlife - a Code of Practice" in conjunction with NCS, SMRU, WWF, MCS, Vincent Wildlife Trust; SSGA, Perth, Scotland. Scottish Wildlife and Countryside Link (1988) "Marine Fishfarming in Scotland - A Discussion Paper", 70pp. Scottish Wildlife and Countryside Link (1990) A Review of Marine Salmon Farming in Scotland", Perth, Scotland, 27pp. Shetland Salmon Farmers Association (1989)"Recommendations for Anti-Predator Measures", Shetland Salmon Farmers Association Environmental Group, Lerwick, Shetland. Thompson, P.M., Pierce, G.J., Hislop, J.R.G., Miller, D. and Diack, J.S.W. (1991), "Winter Foraging by Common Seals (Phoca vitulina) in Relation to Food Availability in the Inner Moray Firth, N.E. Scotland", Journal of Animal Ecology, 60, 283-294.